Confusão conceitual na caracterização da IA: uma análise das percepções e definições estudantis

Abstract

Since generative artificial intelligence (GAI) tools such as ChatGPT has become popular, its use has become increasingly frequent in technological high schools, a phenomenon that challenges traditional educational methods. This study was conducted in the Polo Educativo Tecnológico del Cerro (PTEC) in Montevideo during the year 2024. It focuses on analyzing students' perceptions and definitions, as well as the factors that generate their conceptual misunderstandings when characterizing GAI in terms of its scopes and risks. By digging through virtual discussion forums and interviews, we explore student conceptualizations and the factors that influence them. We discovered that the students tend to associate Artificial Intelligence (AI) with GAI, without considering its historical evolution. This understanding arises from analogies with tools such as search engines, to which they attribute anthropomorphic characteristics. Their understanding derives from their practical experiences with chatbots like ChatGPT, perceived as tools to save time. Furthermore, they acknowledge an ambivalent relationship with it, valuing its usefulness while also recognizing its risks, such as its potential to devalue individual effort. Given the duality between the technology's ease of use and its technical complexity, conceptual confusions become evident. This study highlights the importance of supporting students in building a critical and metacognitive understanding.